Peripheral inflammation and depressed mood independently predict neurocognitive worsening over 12 years.

Background: Neurocognitive (NC) impairment in people with HIV (PWH) is associated with important adverse outcomes, but no markers exist to predict long-term NC decline. We evaluated depressed mood and markers of persistent inflammation, oxidative stress and altered amyloid processing (all common in PWH) as predictors of NC worsening over 12 years. Methods: PWH were enrolled and followed longitudinally in the CNS HIV Antiretroviral Effects Research (CHARTER) study at six US sites. At entry we quantified biomarkers in blood of inflammation: (interleukin-6 [IL-6], C-reactive protein [CRP], monocyte chemoattractant protein type 1 [MCP-1], D-dimer, soluble sCD14 (sCD14), soluble tumor necrosis factor receptor - type II [sTNFR-II], neopterin, and soluble CD40 ligand [sCD40L], oxidative stress (protein carbonyls, 8-oxo-2'-deoxyguanosine [8-oxo-dG]) and altered amyloid processing [amyloid beta (Aβ)-42, soluble amyloid precursor protein-α (sAPPα)] using commercial immunoassays. The Beck Depression Inventory-II (BDI-II) assessed depressed mood at entry. NC decline over 12 years was evaluated using the published and validated summary (global) regression-based change score (sRBCS). A factor analysis reduced dimensionality of the biomarkers. Univariable and multiple regression models tested the relationship between baseline predictors and the outcome of neurocognitive decline. Results: Participants were 191 PWH, 37 (19.4%) women, 46.6% African American, 43.5% non-Hispanic white, 8.83% Hispanic, 15.7% white, 1.6% other; at study entry mean (SD) age 43.6 (8.06) years, estimated duration of HIV infection (median, IQR) 9.82 [4.44, 14.5] years, nadir CD4 104/μL (19,205), current CD4 568/μL (356, 817), and 80.1% had plasma HIV RNA <50 c/mL. Participants were enrolled between 2003 and 2007; median (IQR) duration of follow-up 12.4 [9.69, 16.2] years. Three biomarker factors were identified: Factor (F)1 (IL-6, CRP), F2 (sTNFR-II, neopterin) and F3 (sCD40L, sAPPα). Participants with higher F1, reflecting worse systemic inflammation at baseline, and higher F3, had greater decline in global neurocognition (r ​= ​-0.168, p ​= ​0.0205 and r ​= ​-0.156, p ​= ​0.0309, respectively). Of the F1 components, higher CRP levels were associated with worse decline (r ​= ​-0.154, p ​= ​0.0332), while IL-6 did not (r ​= ​-0.109, p ​= ​0.135). NC change was not significantly related to F2, nor to demographics, nadir and current CD4, viral suppression or baseline NC comorbidity ratings. Individuals with worse depressed mood at entry also experienced more NC decline (r ​= ​-0.1734, p ​= ​0.0006). Together BDI-II (p ​= ​0.0290), F1 (p ​= ​0.0484) and F3 (p ​= ​0.0309) contributed independently to NC decline (p ​= ​0.0028); their interactions were not significant. Neither CRP nor IL-6 correlated significantly with depression. Conclusions: PWH with greater systemic inflammation and more depression at entry had greater NC decline over 12 years. Understanding the basis of this inflammatory state might be particularly important. These findings raise the possibility that targeted anti-inflammatory or antidepressant therapies may help prevent NC worsening in PWH with depression and inflammation.